No abstract

In iterative learning control (ILC), a common assumption is that the initial states in each repetitive operation should be inside a given ball centered at the desired initial states which may be unknown. This assumption is critical to the stability analysis, and the size of the ball will directly affect the final output trajectory tracking errors. In this paper, this assumption is removed by using an initial state learning scheme together with the traditional D-type ILC updating law. Both linear and nonlinear time-varying uncertain systems are investigated. Uniform bounds for the final tracking errors are obtained and these bounds are only dependent on the system uncertainties and disturbances, yet independent of the initial errors. Furthermore, the desired initial states can be identified through learning iterations.

For cold-start recommendation, it is important to rapidly profile new users and generate a good initial set of recommendations through an interview process -users should be queried adaptively in a sequential fashion, and multiple items should be offered for opinion solicitation at each trial. In this work, we propose a novel algorithm that learns to conduct the interview process guided by a decision tree with multiple questions at each split. The splits, represented as sparse weight vectors, are learned through an L1-constrained optimization framework. The users are directed to child nodes according to the inner product of their responses and the corresponding weight vector. More importantly, to account for the variety of responses coming to a node, a linear regressor is learned within each node using all the previously obtained answers as input to predict item ratings. A user study, preliminary but first in its kind in cold-start recommendation, is conducted to explore the efficient number and format of questions being asked in a recommendation survey to minimize user cognitive efforts. Quantitative experimental validations also show that the proposed algorithm outperforms state-of-the-art approaches in terms of both the prediction accuracy and user cognitive efforts.

Variations of deep neural networks such as convolutional neural network (CNN) have been successfully applied to image denoising. The goal is to automatically learn a mapping from a noisy image to a clean image given training data consisting of pairs of noisy and clean images. Most existing CNN models for image denoising have many layers. In such cases, the models involve a large amount of parameters and are computationally expensive to train. In this paper, we develop a dilated residual CNN for Gaussian image denoising. Compared with the recently proposed residual denoiser, our method can achieve comparable performance with less computational cost. Specifically, we enlarge receptive field by adopting dilated convolution in residual network, and the dilation factor is set to a certain value. We utilize appropriate zero padding to make the dimension of the output the same as the input. It has been proven that the expansion of receptive field can boost the CNN performance in image classification, and we further demonstrate that it can also lead to competitive performance for denoising problem. Moreover, we present a formula to calculate receptive field size when dilated convolution is incorporated. Thus, the change of receptive field can be interpreted mathematically. To validate the efficacy of our approach, we conduct extensive experiments for both gray and color image denoising with specific or randomized noise levels. Both of the quantitative measurements and the visual results of denoising are promising comparing with state-of-theart baselines.

Abstract-This paper presents a framework to restore the 2D content printed on documents in the presence of geometric distortion and nonuniform illumination. Compared with text-based document imaging approaches that correct distortion to a level necessary to obtain sufficiently readable text or to facilitate optical character recognition (OCR), our work targets nontextual documents where the original printed content is desired. To achieve this goal, our framework acquires a 3D scan of the document's surface together with a high-resolution image. Conformal mapping is used to rectify geometric distortion by mapping the 3D surface back to a plane while minimizing angular distortion. This conformal "deskewing" assumes no parametric model of the document's surface and is suitable for arbitrary distortions. Illumination correction is performed by using the 3D shape to distinguish content gradient edges from illumination gradient edges in the high-resolution image. Integration is performed using only the content edges to obtain a reflectance image with significantly less illumination artifacts. This approach makes no assumptions about light sources and their positions. The results from the geometric and photometric correction are combined to produce the final output.

Deep convolutional neural networks (CNNs) have shown superior performance on the task of single-label image classification. However, the applicability of CNNs to multi-label images still remains an open problem, mainly because of two reasons. First, each image is usually treated as an inseparable entity and represented as one instance, which mixes the visual information corresponding to different labels. Second, the correlations amongst labels are often overlooked. To address these limitations, we propose a deep multi-modal CNN for multi-instance multi-label image classification, called MMCNN-MIML. By combining CNNs with multi-instance multi-label (MIML) learning, our model represents each image as a bag of instances for image classification and inherits the merits of both CNNs and MIML. In particular, MMCNN-MIML has three main appealing properties: 1) it can automatically generate instance representations for MIML by exploiting the architecture of CNNs; 2) it takes advantage of the label correlations by grouping labels in its later layers; and 3) it incorporates the textual context of label groups to generate multi-modal instances, which are effective in discriminating visually similar objects belonging to different groups. Empirical studies on several benchmark multi-label image data sets show that MMCNN-MIML significantly outperforms the state-of-the-art baselines on multi-label image classification tasks.

Most of the available results on iterative learning control address trajectory tracking problem for systems without time delay. The role of the initial function in tracking performance of iterative learning control for systems with time delay is not yet fully understood. In this paper, asymptotic properties of a conventional learning algorithm are examined for a class of non-linear systems with time delay in the presence of initial function errors. It is shown that a non-zero initial function deviation can cause a lasting tracking error on the entire operation. Impulsive action is one method to eliminate such lasting tracking error but it is not a practical approach. As an alternative, an initial rectifying action is introduced in the learning algorithm. The initial rectifying action is ®nite and used over a speci®ed interval. It is shown to be e ective in the improvement of tracking performance, in particular robustness and uniform convergence. The results are further extended to systems with multiple time delays. An example is given and computer simulations are presented to demonstrate the performance of the proposed approach.

This paper is concerned with the problem of the iterative learning control with current cycle feedback for a class of non-linear systems with well-defined relative degree. The tracking error caused by a non-zero initial shift is detected as extended D-type learning algorithm is applied. The defect is overcome by adding terms including the output error, its derivatives as well as integrals. Asymptotic tracking of the final output to the desired trajectory is guaranteed. As an alternative approach, an initial rectifying action is introduced in the extended D-type learning algorithm and shown effective to achieve the desired trajectory jointed smoothly with a transitional trajectory from the starting position. Also these algorithms with adjustable tracking interval ensure better robustness performance in the presence of initial shifts. Numerical simulation is conducted to demonstrate the theoretical results.